<> "The repository administrator has not yet configured an RDF license."^^ . <> . . "Reconstructing Neural Dynamics Underlying Cognitive Flexibility Using Parameter-Evolving RNNs"^^ . "Understanding the dynamic principles that enable the brain to flexibly adapt behavior in changing environments remains a central challenge in neuroscience. In\r\nthis thesis, I address this question through the lens of dynamical systems reconstruction. I use a reconstruction method, specifically targeted for non-autonomous\r\nneural dynamics from multiple single-unit recordings in the rodent medial prefrontal\r\ncortex (mPFC) during a probabilistic rule-learning task. To this end, I employ a\r\nparameter-evolving piecewise-linear recurrent neural network (pePLRNN), which\r\nexplicitly incorporates time-dependent changes in the underlying dynamical system\r\n(DS). This approach enables the reconstruction of non-autonomous DSs from nonstationary data to characterize of how the neural dynamics evolve across learning.\r\nThe approach was first validated on benchmark systems and task-trained RNNs,\r\nwhere it successfully reconstructed the underlying DS. When trained on the hidden\r\nstate trajectories of RNNs solving artificial rule-learning tasks, the pePLRNN uncovered the dynamic mechanisms by which these networks implemented the learning\r\nprocess.\r\nApplied to electrophysiological recordings from the mPFC of rats, the model successfully reconstructed the non-stationary neural dynamics underlying rule learning.\r\nThe trained model-generated neural trajectories that exhibited the same decoding\r\nproperties as the original data. Change points (CP) detected in model-generated trajectories aligned with those observed in the recorded activity. Simulations of neural\r\ntrajectories under experimental conditions reproduced the behavioral distributions\r\nof animals for both rule types.\r\nAnalyzing the trained pePLRNN as a functional surrogate model revealed that\r\nboth rules were implemented via a single stimulus-dependent attracting region that\r\nguided neural transients toward the correct decision. During learning, this attracting region, along with the trial-specific parameters and latent neural trajectories,\r\nexhibited abrupt changes that preceded the behavioral change point.\r\nThis work establishes a principled framework for reconstructing non-autonomous\r\nDS directly from empirical data and demonstrates how their analysis as surrogate\r\nmodels can reveal dynamic principles underlying the neural computations supporting\r\ncognitive flexibility."^^ . "2025" . . . . . . . "Max Ingo"^^ . "Thurm"^^ . "Max Ingo Thurm"^^ . . . . . . "Reconstructing Neural Dynamics Underlying Cognitive Flexibility Using Parameter-Evolving RNNs (PDF)"^^ . . . "Max_Ingo_Thurm_Thesis_ubpup.pdf"^^ . . . "Reconstructing Neural Dynamics Underlying Cognitive Flexibility Using Parameter-Evolving RNNs (Other)"^^ . . . . . . "indexcodes.txt"^^ . . . "Reconstructing Neural Dynamics Underlying Cognitive Flexibility Using Parameter-Evolving RNNs (Other)"^^ . . . . . . "lightbox.jpg"^^ . . . "Reconstructing Neural Dynamics Underlying Cognitive Flexibility Using Parameter-Evolving RNNs (Other)"^^ . . . . . . "preview.jpg"^^ . . . "Reconstructing Neural Dynamics Underlying Cognitive Flexibility Using Parameter-Evolving RNNs (Other)"^^ . . . . . . "medium.jpg"^^ . . . "Reconstructing Neural Dynamics Underlying Cognitive Flexibility Using Parameter-Evolving RNNs (Other)"^^ . . . . . . "small.jpg"^^ . . "HTML Summary of #37427 \n\nReconstructing Neural Dynamics Underlying Cognitive Flexibility Using Parameter-Evolving RNNs\n\n" . "text/html" . . . "000 Allgemeines, Wissenschaft, Informatik"@de . "000 Generalities, Science"@en . . . "530 Physik"@de . "530 Physics"@en . . . "570 Biowissenschaften, Biologie"@de . "570 Life sciences"@en . .